IT Sustainability Think Tank: How CIOs can navigate the green IT landscape

Data collection

How data is collected will depend on what hosting options are employed. Typically, enterprises have applications hosted in both on-premise datacentres and in hyperscale cloud facilities - and these environments have very different reporting demands. 

In simple terms, reporting electricity consumption from datacentres requires only a sub-meter to distinguish server room consumption from that occurring in attached offices. 

However, the installation of smart power distribution units will provide an understanding of emissions hotspots, which is vital in targeting emissions reduction activities.

Once this data is collected, it needs only a knowledge of the local power generation mix and any company-owned power purchase agreements to convert electricity consumption into carbon emissions. This can be facilitated by some datacentre infrastructure management (DCIM) systems that have these capabilities built in.

A measure of datacentre efficiency that is often used is Power Usage Effectiveness (PUE), which is the ratio of power consumed by computing to total datacentre power consumption and is intended to emphasise the overheads of the air conditioning and other supporting systems.

It should be acknowledged and accepted that reducing consumption from computing alone will have a negative impact on the PUE, but that is not necessarily a bad thing.

Of course, dedicated datacentres require purchasing dedicated hardware. So the emissions from manufacturing that hardware need to be accounted for and – for that - you need to check the supplier’s specifications.

The hyperscale clouds, including Microsoft Azure, Amazon Web Services (AWS) and Google, all have Carbon Footprint Calculators that can paint a picture of the emissions generated by an enterprise's cloud footprint.

Indeed, Google go a step further by providing users with the capability to compare the energy mix on a region-by-region basis at all their locations. Using this can help organisations reduce their emissions simply by hosting in a greener grid. 

Measurement standards

In parallel with identifying what tools to use, enterprises also need to be clear on what measurement standards to report against. The GHG Protocol is the de-facto standard for reporting emissions and is what regulatory reporting requires.

It uses values for the Global Warming Potential (GWP) of different gases (methane, refrigerants from air conditioning, etc.) to calculate carbon dioxide equivalence (CO2e). In this way it allows comparisons between, for example, an on-premise, hardware-based offering with Software-as-a-Service (SaaS).

Data analytics

While regulatory reporting has little requirement for imaginative formatting, it is vital to bring the information to life for team discussions.

This needs the same tools and skills used to visualise data in other areas of the business. Dashboards of graphs, diagrams and tables can help everyone see both what emissions are and the context from which they originate. In this way, mitigation strategies become easier to formulate. Relative sizes can be compared and emissions from less mission-critical services might be addressed more ruthlessly.

While quantifying emissions is necessary for regulatory reporting, as described above, it can also be used to give material benefits by driving emissions reduction activities leading to improved customer perception and cost avoidance.

Measures that can support this include:

• Having a GreenOps dashboard to provide ongoing, detailed reporting of emissions to allow architects and developers to make informed decisions during the implementation of projects
• Managing green tech debt to allow sustainability to be addressed in projects initiated for other reasons
• Allocating the role of sustainability architect to a member of the firm’s architecture review board to ensure projects take sustainability into account in the same way that security is already incorporated into all IT projects now.